What Happens to a Trading Signal After It Is Sent? 90 Days of AlgoVesta Data

To copy signals from a Telegram channel to MetaTrader 5 automatically, start with the Telegram to MT5 bridge.

Data window: 2026-08-03 to 2026-10-31 (90 days, measured on 2026-11-01). Source: AlgoVesta's own production database, queried read-only. Not measured: profit and loss.

How did we measure it?

• Source. AlgoVesta's own production database: the Telegram decision audit, the TradingView webhook log, crypto and MT5 trade records, MT5 login attempts and paper-trading records. Queries ran in a read-only database session; nothing was written.
• Window. The 90 days from 2026-08-03 to 2026-10-31. Most channels launched during the 90 days before, so that earlier window had fewer than five customers per channel and we do not publish comparisons with it.
• Exclusions. Our own synthetic monitoring account (the hourly canary), test-run residue, banned accounts and deleted accounts are removed from every query. In the window this removed 2985 canary and 1067 test webhook rows, 1399 canary crypto trade rows, and 137 webhook, 19 MT5 trade and 57 paper rows linked to test residue.
• Privacy. We publish aggregates alone. Any group with one to four users or events is hidden. No per-customer balances or profits are reported.
• What we did not measure. Profit and loss. We do not publish win rates or returns, and nothing here should be read as a performance claim. Why we do not publish a win rate.

What happens to a Telegram message?

In the window, the Telegram reader received 47288 channel messages for 72 customers across 182 channels.

Most messages in signal channels are not signals: chat, results posts and charts without text. 15.2% of all messages had no recognisable signal structure and 2.8% carried no text at all. The class we care about most is narrower: messages that contain buy/sell, take-profit and stop-loss words but still could not be read. That was 0.2% of messages, across 14 customers.

Of the messages that were stopped before execution, 48.9% were stopped by a rule or setting the customer controls (no bot configured for that account, a channel outside the bot's scope, a daily limit), 48.4% were routine filtering such as duplicates, and 2.0% were attributed to a failure on our side. The full signal path is described in Telegram to MT5 and How AlgoVesta works.

What happens to a TradingView alert?

10755 TradingView webhook alerts reached AlgoVesta from 11 customers.

More than half of alerts arrived while the customer had automatic trading switched off for that strategy. The most common blocked reasons were a disabled webhook (56.5% of blocked alerts) and an alert body we could not parse (27.2%), usually a template with unfilled placeholders.

How long does an MT5 order take?

For MT5 we record the round-trip of each order command: from sending it to the terminal until the terminal confirms.

Our earlier published figure was a single end-to-end demo order at 849 ms. Across thousands of orders the live-server median is slower and the spread is wider. We will keep publishing the distribution rather than a single best case. How orders reach MetaTrader is described in Forex and MT5 automation.

For crypto we could measure webhook receipt to recorded exchange confirmation on the Binance testnet (29 orders, median 4168 ms). Most live crypto orders came through Telegram, where the message-to-order link is not stored, so we have no live crypto timing to publish this quarter.

Did every position have a stop-loss?

The rows cover positions AlgoVesta opened in the window, and "recorded" means the stop-loss or take-profit was present on the position itself. On crypto, the stop-loss is written to our records from the exchange's confirmation alone, not from what was requested. 0 live crypto positions were closed by the stop-loss-failure safety close in the window.

For context, the same MT5 accounts also hold positions that customers opened themselves or through other tools. Those had a stop-loss in 65.4% of cases on live servers and 27.3% on demo servers.

Before execution, Telegram signals were refused 353 times because a stop-loss and take-profit could not be ensured, 541 times because the customer's "require stop-loss" rule applied, and 187 times for invalid stop/target levels. The rules behind these refusals are listed on the security page.

How did positions close?

A stop-loss close is not always a loss: stops are often moved to break-even or trailed. On live crypto accounts, 81.2% of closes were detected on the exchange without a recorded cause (stop, target or manual), which is a gap in our own records that we intend to close.

How long does connecting an MT5 account take?

Of 100 MT5 accounts customers added since we began logging login attempts, 80.0% reached a verified broker login. The median time from adding the account to the initial verified login was 4.0 minutes; the 90th percentile was 27.1 hours. Accounts typically needed 2.5 login attempts (median) before the initial success.

Which accounts are on the platform?

Of active customer MT5 accounts, 56.0% are on live broker servers and 44.0% on demo servers. 98.8% of crypto orders in the window were futures orders. Among active exchange API keys, 19.0% show a permission error as their last error.

How much paper trading happened?

24 customers placed 486 paper trades in the window. We do not publish how many of them moved on to live trading this quarter: the group is below our minimum of five on one side of the ratio.

What are the limitations?

• Testnet and demo are labelled, not blended. Crypto testnet is a shared exchange test environment; MT5 "demo server" is a name-based classification of the broker server.
• Retention gaps. The MT5 login log starts on 2026-08-12, so connection times cover solely accounts added after it.
• Survivorship. Deleted accounts are excluded, and some customers left during the window.
• Selection bias. This is our own customer base, not the market. Customers who choose an automation tool differ from traders in general.
• No performance data. Close reasons are not profits.

What will we measure next quarter?

• Live crypto order timing, once the Telegram message-to-order link is stored.
• The cause of crypto closes detected on the exchange.
• Paper-to-live conversion, if the group is large enough to publish.
• The same tables again, so the series can be compared.

The bottom line

Most of what arrives in a signal channel is not a trade, and most stopped signals are stopped by a rule or setting the customer controls. On MT5, every position AlgoVesta opened in the window carried a stop-loss and a take-profit, and order confirmation took about a second on live servers. We will publish the same tables next quarter so the numbers can be compared.